Y Combinator Seeks Startups to Cut Government Waste with AI-Powered Solutions

Y Combinator is accepting fall applications for its next cohort, focusing on six themes, including cutting government waste using large language models, AI training for blue-collar workers, and building the first 10-person, $100B company. The accelerator wants to fund startups that build software to take over work done by consulting giants like Deloitte and Accenture. Applications close on Monday evening, Pacific time.

Y Combinator, the renowned startup accelerator, is now accepting applications for its fall cohort, with a particular focus on six themes that promise to drive significant innovation. Among these themes, cutting government waste using large language models (LLMs) stands out as a key area of interest, reflecting broader political and economic pressures to reduce wasteful spending.

According to Y Combinator's website, one of the themes for the fall cohort is "Using LLMs Instead Of Government Consulting." This initiative aims to fund startups developing software powered by large language models to replace the work traditionally done by consulting giants like Deloitte and Accenture. The goal is to address the inefficiencies and high costs associated with government consulting, which amounts to over $100 billion annually [1].

Y Combinator's Gustaf Alstromer emphasizes the potential of LLMs, stating that these models are already capable of performing many consulting tasks. However, he believes that startups can do better by building specialized software tailored to government needs. This focus aligns with the mission of the Department of Government Efficiency, a Trump administration initiative aimed at slashing federal spending [1].

In addition to cutting government waste, Y Combinator is also betting on several other themes for the fall cohort. These include:

- AI Training for Blue-Collar Workers: Y Combinator wants to tackle the talent bottleneck in the trades by funding startups that use AI to create personalized training programs. This could help fill the growing shortage of skilled tradespeople needed to build infrastructure behind AI technologies [1].
- Video Generation as a Core Technology: The accelerator is looking for startups that can harness the power of video generation to unlock new applications, developer platforms, and real-time experiences. Y Combinator believes that advancements in video generation models could lead to significant innovations in media, entertainment, and other sectors [1].
- The First 10-Person, $100 Billion Company: Y Combinator is optimistic that small, high-agency teams can build multibillion-dollar companies with just $500,000 in seed funding. The accelerator is focusing on startups that can generate maximum revenue per employee, optimizing for efficiency and execution [1].
- Infrastructure for Multi-Agent Systems: Y Combinator is seeking startups that can develop tools to make multi-agent systems easier to build and maintain. These systems can tackle complex jobs faster by breaking them down and running them in parallel [1].
- AI-Native Enterprise Software: The accelerator is looking for startups that can reinvent enterprise tools from the ground up, with AI at the core. This could include tools for sales, HR, and accounting, similar to how Salesforce and ServiceNow rode the cloud wave 25 years ago [1].

Applications for Y Combinator's fall cohort close on Monday evening, Pacific time. The accelerator is seeking innovative startups that align with these themes and can contribute to significant advancements in their respective fields.

References:
[1] https://www.businessinsider.com/y-combinator-doge-startups-government-waste-ai-2025-8
[2] https://analyticsindiamag.com/ai-news-updates/y-combinator-wants-to-fund-the-first-10-person-100-billion-company/